10039258Optimal Prediction of Sand for AdhesionClosedCollaborative R&DInnovate UKTrain services are affected by seasonal variables particularly leaf fall between September and December. They can also be compromised by wet weather, icy and snowy conditions at a regional or very localised level on a particular route. Maintaining wheel-rail contact to ensure adequate and safe braking requires the use of sand in low adhesion conditions. Sand is dispensed to trains in response to a combination of train service plans and of weather forecast. However, not all trains are currently able to be replenished during overnight stabling and servicing with attendant risks of delays and damage to trains and infrastructure. Also, there is a high level of safety risk when sand replenishment on trains is carried out on a third-rail yard. _"**O**__ptimal **P**rediction of_ **_S_**_and for **A**dhesion" (OPSA)_ lead by Govia Thameslink Railway, the major Train Operating Company on third rail in the UK, will deliver a more efficient and cost-effective means of predicting the dispensing of sand to trains to ensure services are not compromised by adhesion losses and train sets are not required to be removed from planned operating diagrams because of inadequate on board sand supplies. The algorithm developed as a results of this project will base the estimates on an integrated framework that includes the forecast adhesion, track maintenance and the expected speed profile in order to capture the change in weather and the seasonal factors. The algorithm developed represents a cost effective solution to predict the use of sand and schedule the maintenance of trains enhancing in turn safety and reducing the impact of delays on the timetable. The algorithm will be developed including direct measure of sand dispersion, braking, wheel slip and line speed diagram also accounting for human behaviour effects such as driving style. Govia Thameslink Railway has engaged with Cranfield University to deliver the disruptive innovation proposed in this project. The algorithm will enable a more efficient train scheduling improving public performance measure (PPM) addressing train delay targeting in particular the 25% of delay up to 15 minutes cause by several concurrent issues including train rescheduling and the National Rail Passenger Survey satisfaction.